1Paradigms have finite life-spans. Like the generations to which they are tied, they ultimately depart, making way for new explanatory theories. In the best scenarios, new paradigms will prove more relevant than the ones they replace. They will incorporate more variables and will enable a better assessment of the complex reality they seek to interrogate. In other cases, advances made as a result of new paradigms can seem less significant, causing one to question the pertinence of the theoretical innovations put forth. Proponents of hegemonic paradigms, moreover, strongly resist and keenly critique nascent, competing paradigms. But over the long term – on the scale of a succession of generations – the paradigm war and the gradual but unstoppable marginalization of older paradigms yields positive effects. Knowledge slowly becomes more profound, penetrating, and coherent. This occurs despite the fact that in the social sciences, as opposed to what is generally accepted for the natural sciences, a plurality of paradigms is still the order of the day: constellations of competing paradigms succeed one another over time.

2Let us be clear at the outset: several signs indicate that French political science is experiencing a period of paradigm shift. Like any silent long-term transformation, this shift is neither evident to all nor easily discernible. Time must be taken, therefore, to delineate and demonstrate it. New modes of thought have difficulty making themselves visible when up against their predecessors; the presence of the latter is literally blinding. But the present paradigm shift is real, and its importance is immense – so much so that soon it will not be possible to formulate arguments according to previously dominant frameworks. [2] The very title of this article will, before long, become obsolete. An examination of “French political science” will no longer make sense in a future in which paradigm confrontations take place at the international level, as they do in the natural sciences, a point that will be developed below. For reasons of method, it is permissible to refer to “political science in France” – but certainly no longer to “French political science”. The edited volume that has generated the reflections contained in this article illustrates this contention through the provenance of its component essays. The authors are variously German, English, Austrian, French, and Dutch; two of the essays have been translated from English to French, and one from German to French.

3Since the relative stabilization of political science in France – following the period in which the discipline faced the “challenge of professionalization” [3] – the current shift in research agendas [4] is the second to take place. From 1970-1975, French political science was divided among six competing paradigms, most of which subsequently disappeared, while others were profoundly transformed. Twenty years later, from 1990-1995, the outcome was apparent: a different constellation of research agendas now held sway, one that hardly had anything in common with the dominant configuration of 1970-1975. A first paradigm shift had indeed taken place over these two decades. In 2010, an intellectual generation later, the constellation of paradigms is once again undergoing transformation – and the resulting landscape will not resemble the one with which we are now familiar. From this perspective, therefore, I cannot go along with the overly socio-historical analysis of the authors of the conclusion to Rational Choice in Political Science, who assert that since “the paths taken by the various social sciences [are] largely determined by historical and institutional context”, the scholarly debate remains marked by “a certain inertia”. [5] According to these authors, it is thus not possible to observe, “especially in political science”, paradigm shifts such as those described by Thomas Kuhn. On the contrary, I will show that this inertia is relative. With the hindsight of twenty years, the paradigm configurations of French political science are currently experiencing such considerable transformation that it is indeed possible to observe, mutatis mutandis, a “scientific revolution” as articulated by Kuhn. And it goes without saying that in the future, this same process of the incessant replacement of paradigms will continue to occur. [6] Before turning at length to rational choice theories, I will discuss some of the key developments in the succession of paradigmatic revolutions in French political science and describe the underlying social evolutions that provoked such transformations.

4The paradigm configuration of 1970-1975 can be described by drawing upon a report I presented to the French Political Science Association in 1980. [7] Since this report’s cataloging of paradigms was contested neither at the time nor since, I can justify still making reference to it. In brief, at the time, two research agendas were firmly established: an empiricist paradigm – concerning, for example, studies of elections, studies based on polls, or studies of what was generally termed “political life” – and a constitutionalist paradigm – which assigned a determinant role to institutional rules and to central political actors. Two other research agendas were more properly “sociological”, according to the current meaning of the term. An “extensive” sociology (today one would say “historical” sociology) was centered on the longue durée (for example, the work of Stein Rokkan, Barrington Moore, Perry Anderson, and several others), and a new political sociology sought inspiration from sociology’s founding fathers. In addition, a strong Marxist current subsisted in a few sectors of French political science, while still others attempted to promote a research agenda drawn from psychoanalysis.

5Twenty years later, this constellation of paradigms had almost completely disappeared. Beginning in the decade following the 1970-1975 period, three of the six paradigms above became marginal: the constitutionalist, Marxist, and psychoanalytic paradigms. The empiricist research agenda had changed direction, in that the recourse to mathematical formalization and to modeling had been practically banned, both from training in political science and from prominent scholarly production. [8] The historical sociology paradigm, which I referred to at the time as “extensive” sociology, was transformed into something not far from its opposite, the socio-history of politics. The research agendas that were termed “sociological” spread progressively throughout the intellectual universe of political science, while simultaneously splitting into multiple theoretical constructions. At the end of the 1990s, the key themes of French political science could accordingly be listed as follows: interactionism, in its various forms; “critical” sociology, or better, structural-constructivist sociology (Bourdieu); sociology of justification or of “worth” (Boltanski); constructivism, or better, constructivisms; [9] and the socio-history of politics. [10]

6The list no longer ends there. The paradigm shift currently taking place likely does not allow us, at this time, to draw up a new list of the paradigms that are replacing the older ones – a point to which I will return. Yet it is clear that the research agendas of the final decade of the twentieth century have largely lost their scientific efficacy. [11] Consider two paradigms that very recently appeared dominant: first and foremost, the socio-history of politics. It can hardly be contested that this paradigm is today exhausted and incapable of overcoming the two difficulties that prevent it from attaining maturity as a paradigm. The numerous dissertations defended under its auspices do not lead to cumulative knowledge, having so conformed to the French model of doctoral excellence aptly described by Christophe Bouillaud: [12] a singular field (in this case, archival); “a sophisticated empiricism” that does not avoid “descriptivism” [13]; and thus an extreme dispersal of studies that prevents any meaningful pooling of results. Hence the consubstantial impossibility of reconsidering the problems of political science in light of socio-historical studies [14] – with the significant exception of the “mega-object” of European Union construction. [15] There is no real mystery here. Lacking a theory of politics to test, the culture of the historical fact for its own sake prevails, and the production of monographs becomes an end in itself. Advocates of this paradigm thus come together around methodological problems – or to debate their relationship to history, the nature of boundaries between the disciplines, or, according to a famous title, “disciplinary cilices and wanderings”. [16] The most recent synthesis, Pratiques et méthodes de la socio-histoire [Practices and Methods of Socio-History], comes close, in my view, to registering this failure. [17] For my part, I would like to see the return of a historical sociology that, as opposed to the socio-history of politics, seeks to confront major long-term problems. An example of such a study is Jared Diamond’s book on the failure of societies – a decisive issue, if ever there was one. [18]

7A similar conclusion also applies to structural constructivist sociology. This fertile research agenda, which Pierre Bourdieu brought to maturity in the 1980s – and which French political science largely made its own in the era of the “Politix generation” [19] – must likewise be reconsidered. Its political assessments and critical function are firmly rooted in a now-bygone France. That France was still characterized by a traditional political division of labor; and the privatization of classes at both ends of the social spectrum (elites and low-income workers) was still limited. Capitalism had not yet become basically predatory, and financial speculation was not yet computer-assisted. [20] With regard to political science in France, the Nouveau manuel de science politique [The New Political Science Textbook], published in 2009, is like this paradigm’s swansong: a vast work in the service of a paradigm that is disappearing. [21] Those who use this “new textbook” should know that it will hardly give them access to the paradigms at the heart of current scholarly debates. Such access is increasingly provided by other textbooks. [22] The Nouveau manuel de science politique rejects pragmatic sociology, comprehensive sociology, [23] and methodological individualism. [24] Haphazardly, it also rejects interactionism, rational choice, and any formalization or modeling of social facts.

8So while Bourdieu’s sociology is not going to vanish body and soul, it will cease to dominate – indeed monopolize – the field. It will reconstitute itself, trading in some of its fragmented parts for new paradigms, which will no longer apply this sociology faithfully and without reservation. Luc Boltanski, who understands this, has recently undertaken a reintegration of critical sociology and pragmatic sociology [25] – even if he is not the best placed person to do so, having first too unconditionally advocated, and then later too strongly critiqued, Bourdieu’s approach. Young scholars, who were not involved in the genesis of these theories, are the ones who will progressively reconstruct this paradigm, with full intellectual freedom. [26] In this way, a retrospective reconsideration of the sociology of Pierre Bourdieu – now an essential imperative – will take place. [27] One cannot help viewing structural constructivist sociology and the socio-history of politics as perfect illustrations of Imre Lakatos’ remark about “empirically progressive” research agendas (they explore new fields) that are nonetheless “theoretically degenerative” (the research question becomes increasingly impoverished). [28]

9Given the limited scope of this article, I cannot explore at length the social developments that condition paradigm shifts in French political science. Instead, I will present four determining factors: two having to do with the profound changes in the discipline’s scientific practices, and two tied to transformations in French society.

10While still marginal fifteen years ago, [29] the opening of political science in France to the international scene has today become the norm. Young political scientists all work in at least two, if not three, languages. [30] Their scholarship of reference comes far more from the anglo-phone world than from France. This is not a point about the linguistic competence of certain scholars; rather, it is the recognition of a scientific revolution. French political science is finally operating in the manner of the natural sciences. It would be inconceivable for a biologist or a particle physicist to work solely with French sources for lack of English-language skills. To this end, the internet greatly boosts internationalization. Since the mid-1990s, and with an exponential progression, several powerful technological applications have enabled the dissemination of research results worldwide and an increase in the number of potential academic contacts. No longer can anyone invoke the practical impossibility of accessing others’ studies or of making available one’s own. The theories of reference are henceforth the ensemble of theories put forward in the scholarly world. The research questions posed thus become infinitely larger and more varied. It is evidently in this manner that political science, which had become unitary, is now able to be cumulative. [31]

11To take full measure of this development, consider the edited volume whose title evokes this shift of scale, involving political science’s subjects of study as much as its theories of reference: La science politique de l’Union européenne [Political Science of the European Union]. [32] The requirements expected of this book’s contributors are exemplary (even if complied with in variable fashion). Each chapter must begin by providing a discussion of the state of the international field with regard to the subject under consideration, after which it must assess “the way in which French or Francophone political science positions itself in this field”. [33] It must then propose specific avenues of research. In this way, the book elegantly overcomes one of the perils contained in the internationalization of science: the creation of a mainstream, real or imagined, with which the least autonomous scientific communities align. In the blueprint itself for each contribution to La science politique de l’Union européenne is thus inscribed the imperative of an international scholarly reference point. This imperative does not prevent contributors from then opting to use analytical frameworks developed in France – only now, they are equipped with a sound understanding of the field at large. [34] If indeed there is a “new” political science textbook, La science politique de l’Union européenne is surely it.

12The other major development internal to the discipline results from changes in the relative positioning of political science’s various fields of inquiry. For a long time, the scholarly field of reference was the political sociology of states – a field that examined states on the basis of their internal dimensions only. The field of international relations, for its part, remained a domain confined to specialists; it was understood that such scholars would keep to themselves. In this configuration, public policy occupied a somewhat removed place, [35] and comparative studies were rarely undertaken. Consequently, the scholarly theories elaborated in each of these fields remained bounded within disciplinary areas that were still marginal. Rare were the sociologists who thought to question the “realist” or “neo-realist” paradigms in international relations, let alone the “liberal” paradigm. In the 1970s, no political sociologist paid attention to the sequential analysis of public policies or to approaches from administrative science. Fifteen years later, they likewise declined to engage in dialogue with the new institutionalism.

13This ostracizing of theories from dominated fields is no longer possible today. These fields have gained in importance and the debates they produce cannot be neglected. A sociologist of politics cannot ignore what is illuminated by theories of path dependence, advocacy coalition frameworks, or institutional design. Similarly, one cannot call oneself a “constructivist” today without including in this term the specific meaning that the constructivist paradigm takes in the field of international relations. [36] A critical sociologist cannot overlook theories termed “radical” – notably those of hegemony and counter-hegemony – in international relations. Likewise, a proponent of gender theory cannot fail to take into account approaches that demonstrate the shortcomings of an international studies founded on masculinity. [37] Paradigm shifts thus undo the compartmentalization of previously autonomous sub-disciplinary fields. They also reveal the considerable displacement of the boundaries that previously demarcated the various fields of study. Therefore, theoretical investigations in political science can no longer be limited to those stimulated by general sociology. Such a situation attests to a transformation of the world – a transformation that must be understood.

14In effect, paradigms rise or fall in political science for reasons not limited to the dynamics of scholarly debates. They rise or fall also because certain components of political entities are either transformed or disappear. Between the 1970s and today, and even since the more recent 1990s, the world has changed. It has changed, firstly, due to the rearrangements of sovereignty that each state experienced in a globalized – and, on a continental scale, Europeanized – world. While the issue is certainly far from being settled, the debate must be situated at this level. Is a kind of homogenization taking place at the scale of the “planet Earth”, and must we refer to a “global culture”? [38] Does homogenization occur only regionally – for instance, is there a “European model of civilization”, characterized by common elements such as the welfare state, a particular brand of capitalism, compatible party systems, and a stabilized relationship to religion? [39] The possession of political power within each state takes on a different nature when faced with all kinds of transnational configurations. This is not to say that it diminishes, since political power can disappear in one area and increase in another – thus requiring alternative sets of hypotheses. The prodigious development of tools of mass communication, via the interface of computers, progressively modifies aspects of social relations between individuals, between firms and clients, between industrial and financial actors, between readers and print media, and between political parties and their members. To understand these transformations, new research strategies – or at least expanded ones – are needed.

15The transformations experienced by states, which can no longer be understood solely as distinct national entities, are accompanied by social transformations equally fraught with consequences. In a country like France, the scope of these transformations is prodigious, involving a collapse of religious practice; a massive increase in the number of non-traditional families and thus a transformation of parental roles; a lengthening of the average life expectancy; an increase in permissiveness and a weakening of social control; a lowering of the age at which young people become independent; a “digitization” of the world through the widespread use of the computer and the smartphone; etc. Readers can complete this list as they wish – a list whose banality tends to make us take for granted the formidable social transformation it represents, one that is affirmed by all comparative research on successive generations.

16Consider, therefore, whether this transformation reflects, rather than produces, an evolution in which man is more and more focused on himself and his immediate circle. This observation applies to criminality: because the capacity of man for self-control, for the repression of desires, and for the respect of prohibitions has been modified, the use of violence has necessarily become more common. [40] It also applies to individuals’ principles for action: if man has become more individualistic, calculating, and opportunistic, he will now be more inclined to compare the costs and benefits of his choices, and to become a “rational actor”. Here the connection between social change and sociological paradigms becomes apparent, and the authors of Rational Choice in Political Science do not fail to emphasize this link, at least as a point to be debated. If man today has indeed become the “selfish and materialistic individual” produced by social evolution, then rational choice theory becomes a “particularly powerful and useful explanation of social life”. [41]

17At the time of this article’s writing, the paradigm shift in political science in France has not yet fully taken place. It will become more visible in the years ahead. For now, I can still attempt to predict and enumerate the paradigms that will soon become dominant – or if not dominant, that will at least become prerequisites in a political scientist’s training. These include new research agendas originating in constructivism and a paradigm resulting from the renaissance of historical macro-sociology. They also include a new integrative sociology, based on more extensive research, which has reconsidered its epistemological status and come to terms with its empirical dimension. Finally, they involve paradigms of gender, of pragmatism, of a renewed interactionism, and – here we get to the hub of the matter – of rational choice theory. The authors of Rational Choice in Political Science are fully aware of inserting themselves into the succession of paradigms discussed above. In this volume, therefore, one finds a study of the effects of the succession of scholarly generations in France [42] as well as an analysis of the shift in Germany from a paradigm of regulation to one of governance. [43] From a perspective that ranges across the discipline, the volume also contains reflections on the causes and manifestations of the considerable divide between the United States and Europe with regard to the implementation of rational choice theory. [44] I turn now to an examination of this theory.

18* * *

19Readers of Rational Choice in Political Science will first seek clarification of the meaning of rational choice theory. They will find such clarification throughout the book, in different forms. The contributors to this volume frequently provide their own definitions, and it is instructive to critically compare them. Consider a few examples. For Olivier Pamp, the rational choice approach, “drawing on game theory, contents itself with advancing hypotheses about actors’ decisions so as to model their strategic interactions”. [45] Likewise, Aurélien Evrard remarks that the rational choice approach “aims to highlight the role of actors, whose preferences are seen as clearly identifiable and stable, and who develop particular strategies in order to inaugurate or to oppose a change”. [46] In a manner that is both more concise and more precise, Pieter W. G. Bots explains that the notion of rational choice involves what I will call three indispensable elements (Bots calls them “concepts”): “an actor capable of choosing, a series of options among which he can choose, and the capacity to reason so as to select an option”. [47] He then immediately specifies, as if it were obvious, two points that in fact provoke debate: that an actor’s ability to reason implies a “capacity to make logically correct inferences”, and that “the actor is capable of evaluating the ultimate consequences of his choice of a particular option”. [48]

20In general, these definitions – which occasionally appear in other guises elsewhere in the book – are only preliminary. A different and much more useful basis for discussion is provided in the book’s introduction and conclusion, in the form of a list of postulates of rational choice theory. Borrowed in part from Raymond Boudon, these postulates, in abridged form, are as follows:

21

“1. The fundamental unit of society is the individual (postulate of methodological individualism).
2. Preferences are a given. One does not attempt to understand the reasons for actors’ preferences.
3. Preferences are ordered and transitive.
4. The agent undertakes a cost-benefit calculation with the aim of optimization/maximization.
5. In order to undertake this calculation, the agent needs information about his possible choices, the consequences of his actions, etc.” [49]

22Each sentence, and almost each word, in such a list raises sociological problems – which cannot all be addressed in this article. Instead, I will consider here the problems that are presented as “critical debates” by the book under review. It is in these instances that the book offers the most avenues of inquiry and provides the most exemplary demonstrations. The first issue concerns the meaning of rationality for a theory that describes itself as “rational” in its very title. The second is about the universality of the theory or lack thereof. Does it aim to apply to all social processes (since, as the first postulate above states, society is comprised of individuals – and we know that individuals are constantly making choices) or only to some social processes? Or does it only apply at certain specific moments? The final issue involves rational choice theory’s methods of analysis: precisely because they are “rational”, can we not – and must we not – assess them using the more effective scientific tools provided by mathematics? By allowing formalization and modeling, rational choice theory would thus enable the social sciences to become more rigorous and fertile. I turn now to an examination of these three issues.

23If one adheres to the five postulates presented as an initial reference point in Rational Choice in Political Science, a matter of central importance is eclipsed. This becomes apparent if we consider the somewhat cynical formulation, advanced in 1975 by François Bourricaud, of the five axioms of “neo-individualism” (the term he used at the time). With admirable dryness, Bourricaud laid out these axioms as follows:

24

“1. To explain the functioning of a society, the preferences of the individuals that comprise it are relevant.
2. The individual actor makes choices according to a subjective assessment of his interests.
3. The individual actor is indifferent – or at least does not give the highest priority – to the consequences that his action can bring about for others.
4. The individual actor ignores – or at least only vaguely grasps – the consequences that his decision can have for others.
5. The individual actor is incapable of predicting, and thus controlling, all the consequences of his own action.” [50]

25A superficial reading would lead one to think that these two lists of five points each come from the same intellectual universe. Yet this is not the case. As a proper methodological individualist, Bourricaud was a non-consequentialist: the individual actor could not know the consequences of his decisions for others – even more so because he was already largely incapable of predicting the consequences of his decisions for himself (and, as a general rule, knowing the consequences for others would in any case leave him indifferent). Conversely, rational choice theory is necessarily consequentialist: when making a choice, the actor habitually assesses the consequences that can result for himself. Rational choice theory – and this is a crucial element – also holds that the actor can, at least in part, effectively predict the consequences of his decisions. In some cases, moreover, the actor takes into account the consequences for others of his decisions – at least insofar as such consequences could ultimately be unfavorable for him. Assessment of such consequences thus involves several different kinds of judgment: axiological, instrumental, probabilistic, etc. This point opens up a considerable epistemological and philosophical debate, one that I cannot delve into here, especially since it is virtually absent from the book under review. [51] But the existence of such a debate must not be forgotten; indeed, it will be recalled many times in the course of the following analysis.

Rational choice: rational in what sense?

26A first “critical debate” is opened over the issue of the content that rational choice theory gives to the “rationality” of the individual. At first glance, one would be tempted to retain the most typical meaning of the term “rationality” – that which characterizes homo economicus. Qualifiers abound: “selfish”, “materialistic”, “utilitarian”, “individualistic”, “anti-social”, “calculating”, “strategic”. [52] But this would be to close the debate before it is even opened. Several scholars rightly explain that the rational choice approach can easily account for “apparently irrational behavior”, since such behavior can be compared to the behavior that a rational attitude would produce in the same situation. As Bots remarks, “An actor can be poorly informed (some of his factual beliefs are false), mistaken (some of his causal beliefs are false), or misled (some of his evaluative beliefs are false).” [53] I will clarify this observation with three remarks.

27Firstly, no one contests that the choice of ends belongs to the individual; and that when analyzing an action sequence, these ends must be considered as data. The individual can rationally pursue objectives that are not in the least bit selfish or materialistic. He could wish to devote himself to a collective cause or to make sacrifices for a good that he deems superior – all while fully remaining a rational actor who is able to calculate how to achieve these preferences. As Pamp remarks, “Contrary to an overly widespread idea, actors’ preferences are not necessarily materialistic in the strict sense. A utility function can encompass all kinds of altruistic or social ends.” [54] To say, therefore, that the rational choice paradigm “does not offer very heartening postulates about human nature”, or “is hardly comforting” is to aim at the wrong target. [55] In fact, in my view, the opposite is the case – as will be explored later in this article.

28Secondly, one must take full measure of the meaning of a “utility calculation” in social life. It can certainly refer to a pure Olsonian calculation, with which everyone is familiar – that which leads the calculating worker to hope that a strike will be massive so that it extracts concessions from his employer and to decide not to participate in the strike so as to benefit from a “free pass”. But rational calculations multiply rapidly, forcing the actor to make far more subtle decisions that considerably complicate the analysis but do not escape the model. Yves Schemeil appropriately cites Erving Goffman and Howard Becker, “whose actors consistently make more or less successful calculations”. He notes, furthermore, that “Goffman’s work abounds with episodes in which the actor’s sagacity is severely put to the test, compelling him to make incessant calculations about the utility of his most habitual gestures.” [56] For instance: I had decided to go to the demonstration, but it’s starting to rain; should I still go? Or: I had decided to go to the demonstration, but then a sick friend called me to ask a favor; should I give priority to my friend? Or: I had decided to go to the demonstration, but then I learned that a union I really don’t like will lead the protest march; should I attend or not? [57] In the same way, the vote has been the subject of very detailed analyses that seek to understand how the voter makes a choice. For the past two decades, scholars have thus explored not the rational voter, as conceived by Downs, but rather the promising line of inquiry of the “reasoning voter”. The latter approach has enabled a more satisfactory understanding of the rationality of voting. [58]

29Finally, even the “absurd decisions” studied effectively by Christian Morel [59] can be elucidated through rational choice theory. To highlight cognitive errors by analyzing extreme cases – in which decisions made have consequences that are exactly the opposite of those intended, sometimes tragically so – is not to discard the rational choice paradigm. In such cases, rationality is actually constantly present, at three levels: that of the analyst exploring “radical and persistent errors” in a perfectly rational manner; that of the social actor being studied precisely for his flawed, or more often suspended, [60] rationality; and that of a rationality whose proper use would have avoided the catastrophic consequences of the errors committed. I will subsequently return to these three aspects of rationality.

30At its core, rational choice theory implies only that the social actor has the capacity to set goals for himself, that he is able to decide what he is or is not going to do, and that he makes decisions following a process of reasoning – a calculation about what seems preferable to him. Viewed thus in elemental form, rational choice theory is a tool for describing an observable sequence that can be logically reconstructed. It does not mean, or at least it must not mean, that such a description of a decision sequence, and often of a strategic interaction, exhausts the explanation sought. Everyone knows that a young high school graduate makes a choice about which field of study to pursue in college based on pressing social factors. Yet it cannot be denied that this same graduate sets goals, obtains information, undertakes a process of reasoning, and makes a decision. I do not see why a sociologist would thus close their eyes to this and depict the actor as a sleepwalker, unconscious of the forces that drive him. To restate the obvious, the young graduate’s decision might be unexpected, even surprising, given what the social factors would seem to dictate. Moreover, the combination and complexity of these factors prevent us from predicting with a minimum of certainty the decision the student will take. Finally, the considerable expansion of higher education options in France over the past several years allows for far more individualized strategies than in the past. What probably distorts the debate over rational choice theory is left unsaid: when there is a choice, it is assumed that there is freedom; and sociologists are loath, with good reason, to describe the social actor as “free”. Rational choice theory is thus rejected because the actor’s freedom of choice is rejected. I have treated this problem at length elsewhere. [61] Here, I will just say that the concepts of freedom and social determination exist in two intellectual universes that do not meet and that cannot semantically meet. All sociologists are led to refer to man’s “power to act” in objectifying terms (autonomy of the intentional agent, social practice produced by schemes of perception and evaluation, capacity of an actor to choose between several options, etc.) – which exempts them from the impossible investigation into human freedom.

31Studies based on rational choice theory constitute just one particular case. The essential characteristic of all scholarly studies, and not only in sociology, is to limit, via a methodological decision, the part of the real on which the inquiry focuses. This is done, however, without ignoring that other parts of the real exist and can also be elucidated. The least scholars can do is grant a presumption of scientific rigor to works that examine the parts of the real that we ourselves do not study – and thus not reject outright paradigms constituted to deal with these other parts. The editors of Rational Choice in Political Science are certainly tempted to stop there. According to them, the rationalist approach must be envisioned “as a particular and partial mode of deciphering social phenomena”. It is, in effect, “a simple tool for knowledge”. [62] A final slip of the tongue perhaps reveals the core of their thinking: they call for a “combination of rational choice theory and of approaches that are more sociological”. [63] They thus place rational choice theory on the side of approaches that are not very sociological in nature – leading one to wonder whether they should have been bolder in their stance.

Rational choice: but only sometimes and in certain areas?

32The evidently central question of rational choice theory’s sphere of validity opens up a second particularly rich “critical debate”, one that concerns the domains that lend themselves more than others to use of the paradigm. Here it is necessary to return briefly to the fundamental distinction between methodological individualism and rational choice theory. For methodological individualism, the social world is a sum of individuals. To be studied sociologically, all the social entities with which one is familiar (the state, institutions, systems of representation) must, in the final analysis, be boiled down to the individual “atoms” of which they are constituted. For its part, rational choice theory does not adopt this vision of the social world. Even if it can be applied to individual subjects – for example, the prisoner faced with his dilemma – rational choice theory does not assume methodological individualism in the way that, for demonstrative purposes, game theory does. The paradigm can just as well apply to groups, and rational choices are present in political or union action, in companies, and in institutions of all kinds. In this sense, the state itself is, or can be, a “rational actor”. Rational Choice in Political Science explores three ways of addressing this key issue of a privileged sphere for the application of rational choice theory. I turn now to an examination of these three approaches.

33The first approach consists of designating research fields in which the rational choice paradigm seems like it should naturally apply, given that the decision-making processes at hand aim to be as rational as possible. This is the case in economics, for companies and consumers alike – and even more so for investors, whose decisions, as noted above, are based on advanced mathematical formalizations. [64] This is also the particular case of European studies. Scholars in this field remark that they do not expect to see it suddenly become a notable playground for rational choice theory; instead, European studies seems to offer a “space” suitable for a “supple variant” of rational choice theory. [65] Processes of Europeanization, in effect, presuppose the following: actors, hierarchies, actors’ preferences, interactions, strategies, action sequences, institutions and the procedures that they establish, informal rules, and recognized “ways of doing things”. [66] It is clear that all these terms would seem to encourage the use of rational choice analysis, in the broad sense that I have given it. And in fact, one finds these elements – beyond a “theoretical and methodological pluralism” that now is surely bypassed – in studies on the following subjects: negotiations within the European Council, the dynamics involved in European Union directives, cross-polity coordination (for example, in higher education or in defense), the adaptation and representation of interests at the European level, and the increasing socialization of national political actors in carrying out projects financed by the European Commission. Beyond the phenomenological variety of these fields, the research conducted always tracks interactions between agents with clearly identified objectives and obvious hierarchies of preferences. These agents know how to make use of the rules of the game, to locate strategic actors, and to carry an action through the medium term. They are, in short, actors who know how to make rational choices.

34The rationalist paradigm is also prevalent, for several scholars, in the domain of international relations – where, for example, negotiations between states are conducted by upper-level civil servants who work collectively to assess as precisely as possible the consequences of such negotiations. The mastery of inter-governmental exchanges is such that Gerald Schneider can assert that the “postulate of absolute rationality seems defensible with regard to negotiations between governments”. [67] This postulate can be expanded to include the unilateral decisions of states in international affairs – or at least those decisions that so engage the state (such as declarations of war) that they are made after intense discussions among the highest-level political leaders of the moment. [68] Even in dictatorships, where decisions of this nature are ultimately made by the dictator himself, they always involve discussions with military leaders, close advisers, and government or party figures. Such discussions enable scholars to get extremely close to the final solitary decision-making.

35A reading of one of Ian Kershaw’s recent books, Fateful Choices: Ten Decisions That Changed the World, 1940-1941, [69] allows us to give full meaning – if perhaps not the intended meaning – to Schneider’s postulate. In effect, Kershaw studies in meticulous detail how these fateful decisions of 1940-1941 were made. They include, for example, the English decision of May 1940 to refuse any negotiation with Germany and to pursue war; Hitler’s decision to attack the Soviet Union; and Roosevelt’s decision to break from American neutrality and from the isolationism dominating American public opinion in order to aid Britain. Consider, for a moment, Japan’s decision, made in the fall of 1941, to wage war against the United States. This decision was achieved with the surprise attack on Pearl Harbor of 7 December 1941. It was the subject of extraordinarily long discussions at almost daily meetings over the course of three months. The meeting of 1 November 1941, for instance, lasted seventeen hours. The proliferation and duration of the debates do not necessarily correspond to the rationality of the final choice made. Yet they do at least attest to the fact that arguments were weighed, that all possibilities were very closely examined, that consequences were scrutinized, and that the reasoning of the various parties was tested – made to wrestle against other rationales and organized progressively into a hierarchy. Along the way, moreover, the Prime Minister, Prince Konoe, was driven to resign. He was replaced by the Army Minister, General Tojo – who shared, for a time, some of the reservations that had led his predecessor to resign. All things considered, was the decision to attack the United States a “rational choice”? Evidently not – if we hold to the classic canons of rationality – since the Japanese decision-makers finally agreed that defeat was inevitable in a war with the United States, and that waiting to fight this war (lost before it began!) would result in an even more disastrous, and thus more immediate, defeat. [70] The rationality at work was of another nature. The Japanese leaders could not accept a “loss of national prestige”. They believed they had attained prestige with their recent imperialistic expansion. They wished to defend to the end their “national honor” and to avoid becoming a third-rate nation. Prime Minister Tojo expressed this stance in lapidary fashion: “If this war was for self-existence, then we must be prepared to wage war, even if we foresaw eventual defeat.” [71]

36This case study, although too briefly explored here, contains rich lessons for our discussion. During those three months in Japan, it was indeed a “rational choice” that was undertaken, in the sense of the rational choice paradigm. All positions were expressed, information about the situation was weighed, consequences were evaluated, arbitration took place after in-depth debates, and a decision finally emerged. This rationality, however, was that of a social group making a decision. The group could rationally, for example, prefer defeat to dishonor (perceived as such). It could, as noted above, take the risk of a “national hara-kiri” rather than submit to foreign dictates. [72] Here, we again encounter the central issues previously evoked. We cannot, as easily as rational choice theory suggests, put aside the rationality of the reasoning undertaken to make a choice, the preferences outlined, or thus the subjective assessment of what is decisive (honor or defeat?). In the final analysis, a rational choice is a choice that can be explained by an observer, himself equipped with scientific reason. It is in this sense that Kershaw’s Fateful Choices can be taken as an example of the use of rational choice theory – even if it is not only that. [73] If an observer is able to explain a choice, it is because this choice was, at least in part, conceived by the actors themselves as explainable. The actors themselves understood their reasons, and thus, their rationality.

37The second method for allowing rational choice theory a sphere of significance is to consider it as one possible approach among others within an explanatory system that is eclectic, or at least agglutinative. The classic example of this, which unfortunately provokes despair, is that of the “three new institutionalisms”. [74] Rational Choice in Political Science accords this position another opportunity to express itself, by means of a long article by Mark Aspinwell and Gerald Schneider. [75] This article was first published in the European Journal of Political Research [76] and then translated into French – as was the famous, and equally despair-provoking, article by Peter Hall and Rosemary Taylor. [77] Aspinwall and Schneider insist upon placing at least two approaches under the same umbrella of “new institutionalism”, which they then study simultaneously: sociological and historical, or constructivist, institutionalism and the institutionalism of rational choice. They pursue this method even while stressing that these two approaches are opposed and that, moreover, “no consensus definition of institutions and their salient attributes has been reached”. [78] For some scholars, institutions are “modes of rational behavior”. For others, they are “coherent and long-term repertories of thought or of behavior”. For still others, institutions are “shared symbolic universes”, or even “institutionalized culture”. Formal rules, procedures, norms, and cognitive frameworks can also be included; as can ideas, representations, and “the subconscious and pre-rational dimensions of human behavior”. Finally, “institutions” can refer to short-term decision-making strategies as well as to long-term processes of internalization or myth-making. One could write a “problematized history” of the term. [79]

38When “institution” and “the institutional” reveal such extreme diversity, it has to be recognized that the term loses its distinctive meaning and can be used in any and every way. It is a mystery of the social sciences that such high-quality minds have devoted so much time, energy, and ink attempting to convince us that the three or four new institutionalisms have something in common – despite consistently demonstrating that these institutionalisms are opposed to each other in almost every respect. If a regulatory authority for political science existed, one of its first actions would have to be to outlaw the expression “new institutionalism”. The term “institution” has lost a stable meaning, and all sociological studies dissolve into utter chaos with regard to it. After 36 pages of effort, this is essentially the conclusion that Aspinwall and Schneider reach when they remark, “From all perspectives, the future of institutionalism in European integration studies thus looks daunting.” [80]

39I gladly shift gears now to address the “model of the three I’s” put forward since the late 1990s by Yves Surel and Bruno Palier [81] and included in Rational Choice in Political Science in the form of a long article on the topic by Surel. [82] Within the limited framework of this review essay, the model of the “three I’s” calls for three remarks. Firstly, while Surel explicitly situates himself as an intellectual descendant of the three new institutionalisms, he also observes that “the varied definitions of the notion of institution” in this theoretical haze are “detrimental because they serve as a source of confusion”. The “three I’s” thus allow for better “distinctions” to be made. Moreover, they even enable a “transcending” of the various institutionalisms – by proposing that scholars work from interests, institutions, and ideas, on the basis of clear analytical definitions.

40Secondly, Surel cautiously adheres, for the time being, to an approach that he describes as “methodological and heuristic”. He defines this approach as “a research method with a descriptive quality” and as a “protocol” that can help to “define combinations of analytical elements”. There is nothing objectionable in this – even if it is surely too limited an ambition. Its most marked contribution lies in the sequencing operation that the model calls for. This sequencing takes place at two levels. Each of the “I’s” has its own temporality, and each study must reconstruct these polyphonic temporalities. In addition, in any process, the role of each variable differs over time. Interests or ideas can be determinant at any given moment, since the weight of institutionalization dominates the sequence until interests once more take over. In this way, the model of the “three I’s” offers its solution to the issue of a legitimate domain for rational choice theory. The rationality with which social actors are endowed (Surel specifies right away that this rationality is “limited”) is taken into consideration at certain moments, which must be located. Analysis of these moments is then conducted via the rational choice paradigm – which includes “the use of formal and quantifiable indicators”. [83] This approach still entails, as Bouillaud rightly observes, a simple organization of “multiple interpretive frameworks embedded within each other”. [84] Yet the pedagogical clarity, the honesty of the enterprise, and – it must be recognized – the mnemonic cleverness of its naming deserve consideration, especially since the “three I’s” open, in their way, the investigation I now undertake.

41French political science seems to be in the process of balkanization. To paraphrase Bouillaud, it seems to be engaged in the endless fragmentation of its object of study. [85] This development is certainly consistent with the dynamic taking place in all sciences, by which the apportioning of that part of the real which is to be studied is becoming increasingly narrow. This narrowing has as much to do with the sociological characteristics that govern scholarly communities as with the nature of research. Yet it must not be forgotten that in the natural sciences, this fragmentation occurs alongside a quest for a general theory – such as the “Grand Unified Theory” of fundamental interactions in physics. Such a quest is cruelly lacking in political studies today. [86]

42For those who would set out to construct such an essential general theory of politics, the questions to be taken into account are neither surprising nor unsolvable – yet it is necessary to know how to articulate them. Reference to rationality has its place at several levels. What is the role of rationality in the social construction of politics? [87] And what contribution today do the results obtained by social science research make to the social construction of politics? How does the rationality involved coalesce with symbolic, aesthetic, ethical, and even sentimental [88] modalities to constitute various entities’ relationships to politics? With regard to political action, what is the role of rational calculation in individual decisions (to vote, to join an organization, to mobilize, to emigrate) and in collective or institutional decisions (to judge an alleged offender, to negotiate at an international conference, to select the layout for a high-speed route or for a pipeline)? It must be recognized that rationality is at least partially at work in these situations; and therefore, that rational choice theory has some degree of sociological validity – especially since the “rational phase”, in the sense that I am giving it, exists in practically all cases. [89]

43But many other social and societal mechanisms must be incorporated into the “grand unification” of politics. I can only allude to these mechanisms here. What of, therefore, that which affects the totality of the political edifice: that which constructs or destroys the social order; that which produces breaks in temporality (a revolution or a crisis); that which weds together, depending on the moment, previously distinct social trends in the form of the mathematical image of strange attractors; and that which adheres to the conjuncture upon which the bulk of interaction theories focus. And what of the macrostructural basis upon which political society (to use this mediocre expression) is built: a territory, the social groups living within it, the gendered division maintained there, the social fields established there, the language or languages spoken there, and the relationships with the world developed there.

44After a long period in which theory centered on the conditions for valid sociological knowledge, the time has perhaps come to focus on the objects of this knowledge. This would involve exploration of the ways in which the great political mysteries and the discipline’s foremost areas of inquiry are organized. Is it not time to have at our disposal a sort of “periodic table of the political elements”? This would enable us not only to fix the list of these elements, but – to extend the metaphor of Mendeleev’s classification – to describe the forces of attraction between these elements, the diagonal relationships that link them to each other, the energies put into play, the regroupings to be undertaken, and the resulting properties. Rational choice would occupy several squares in this periodic table – a block, in fact – and it would make connections that varied according to the table’s rows and columns.

Rational choice: because it is rational, can it be formalized?

45The third and final critical debate opened up by Rational Choice in Political Science goes beyond the rational choice paradigm to examine the attitude of French political scientists towards modeling, quantification, and formalization in their discipline. Bouillaud’s contribution to the edited volume, which I have frequently cited, is devoted to this subject. Bouillaud demonstrates, firstly, as if it were necessary, that mathematical formalization has been virtually absent from French political science in recent decades. He shows that, in fact, an “aversion” for the practice has increased. He then puts forth a series of six explanatory and very convincing hypotheses; here, I will content myself with highly recommending consultation of these hypotheses. To consider the rational choice paradigm in this context potentially allows French political science to remove a few obstacles that compromise its scientific soundness. Formalization and modeling oblige the user to define the selected variables in a comprehensive manner, for purposes of their operationalization. Several of the authors insist on this point. As Yves Schemeil contends:

46

“With the rational choice method, the protocols established in the sciences are taken seriously: define one’s object, model it so as to be able to experiment with variations in the testing conditions, observe frequencies and occurrences. To undertake modeling […] is to reduce the number of variables and to take bets on the direction of causality.” [90]

47For his part, Olivier Pamp asserts:

48

“Formalization enables clarification of the hypotheses with which one begins. It allows one to deduce the logical implications of these hypotheses and to formulate clear predictions that can be tested quantitatively. Empirical tests subsequently offer the possibility of refining the model’s parameters.” [91]

49Finally, Pieter W. G. Bots makes the pointed remark that the use of a model [92] “requires the analyst to make explicit, with more precision than in other methods, his own thinking about the modes of reasoning employed by the various actors he is studying”. [93] In effect, the least one can expect from a scientist is that he be compelled “from the start, to reflect on his own assumptions – in other words, that he be obliged to engage in a reflexive exercise”. [94] If scholars must be reminded of this point, it is undoubtedly because their daily practice, in which formalization is absent, has moved away from this imperative.

50The anticipated result of modeling is to rigorously establish conclusions that could resemble scientific “laws” and thus be subjected to critique; and it is also to embark on the process of cumulativity that is so cruelly lacking in a large swath of French political science today. I will take the time here to provide a detailed example in which the “laws” (or, more modestly, the “hypotheses”) should stimulate discussion, confirmation, or refutation. Pamp’s essay in Rational Choice in Political Science addresses European Union member countries’ policies to restrict public spending in the period prior to the introduction of the euro, 1990-2001. [95] He formulates a series of hypotheses from the following simple initial observation: “people with low or average incomes prefer a situation of low unemployment generated by high growth, and they accept its correlate, relatively high inflation” – all the more so because “inflation objectively benefits low income groups, in that it tends to flatten income distribution”. He continues, “Conversely, property owners and those with above-average incomes prefer low inflation, even if this means higher unemployment.” [96] On this basis, an elementary model of economic policy can be formulated. As Pamp outlines:

51

“If a left-wing party wins an election, it will seek, in accordance with the preferences of its voters, to lower the unemployment rate; this policy will result in an increase in growth and in higher inflation. If a right-wing party wins an election, it will take action, in accordance with the preferences of its voters, so as to lower inflation. The result will be a recession, during which unemployment will rise and inflation will decline.” [97]

52Refining the model, however, causes counter-intuitive elements to crop up and the banality of the initial points to fade away. As Pamp elaborates:

53

“A foresighted and rational right-wing government can anticipate an electoral defeat. To the extent that such a government dislikes the public policies that its left-wing successor would put into place, it can choose to accrue debt. In this way, it forces the future government to dedicate resources to the debt at the expense of the latter’s preferred objectives – which include, for instance, reinvigorating the economy or increasing social services expenditures.” [98]

54This provides new insight into the economic policy pursued since 2007 under Nicolas Sarkozy’s presidency. The left can evidently do likewise; it can reduce deficits so as to cause its right-wing successor government to use surpluses by, for example, reducing taxes. Another lesson from game theory is pertinent here, one that has to do with coalition governments, according to the “war of attrition” model. As Pamp explains:

55

“Each party in the coalition has an interest in placing the burden of fiscal adjustment on the other parties’ voters – which incites coalition members to impede solutions by waiting for others to take the initiative. This model thus predicts that a lowering of deficits is improbable, even if everyone agrees on the fact that it is necessary.” [99]

56A new “law” thus follows:

57

“As the number of parties in a coalition government increases, a policy of decreasing public spending becomes less likely.” [100]

58Successive hypotheses must evidently be combined, so as to enable a more precise specification of the sphere of validity of each hypothesis. If only a few parties are associated with the coalition government – and even more so if a single party is in power – they will “strategically use the debt to constrain future governments to act in accordance with their own partisan preferences”. [101] Even the left is subject to this rule: a government of the left that is not a coalition government will be driven to carry out a policy of fiscal restraint. Recognize that this result veers off the beaten path of the sociology of public policy – and that it merits attention for its failure to take reality into account. Conversely, when power is fragmented, parties behave as the initial law stipulated: the left spends freely and the right is more restrained. The reader will forgive the length of this exposition, which is intended to make the case for the relevance and usefulness of modeling as an approach in political science. I at least maintain hope that such an idea will still be heard. In addition to the one provided by Pamp, the book under review contains several other examples. [102] I also recommend other studies that explore this approach – an approach that will ultimately assert itself. [103]

59To end this section, allow me to make a personal comment. I cannot help but feel great sadness as I observe the striking regression in this domain of the dominant version of present-day French political science. Like others of my generation, I started teaching at a time when one notably had students learn the “laws” of parties and of party systems as established by Maurice Duverger. One could discuss these laws at leisure, refining or contesting them on the basis of works by Douglas Rae, Jean Blondel, Giovanni Sartori, Jean Charlot, and Georges Lavau. No one feared typologies or statistical constructions, which could turn out to be heuristic; no one objected to broad international comparisons; no one evaded the search for “laws” of social functioning; and no one thought it unseemly to test the conclusions of earlier scholars and to make one’s own contribution to a shared intellectual structure. At that time, one would have imagined that, 30 years later, we would have available to us an extensive scientific corpus – and that numerous regularities would have been observed and formalized in several political science fields. Today, unfortunately, scholars no longer seek “laws” of social functioning. This is due to the balkanization of the discipline, [104] to exacerbated forms of competition for academic jobs, to the general loss in our discipline of any mathematical or simple statistical competence, and to the obstinate refusal to conceive of the natural and social sciences as still linked by a kinship of methods and objectives [105] – in addition to a few other reasons of which everyone is aware. “Laws” of social functioning simultaneously make political society more intelligible and constitute the pedagogical base upon which a learned society is built.

60* * *

61The volume edited by Mathias Delori, Delphine Deschaux-Beaume, and Sabine Saurugger is thus immensely relevant because it raises, directly or indirectly, three issues that are central for the future of contemporary political science.

62In essence, underlying any use of the rational choice approach is the affirmation that the reasons behind a decision, an action, a behavior, or a preference can be elucidated. With the exception of cases in which man makes a decision by drawing lots (and even then, one could reconstruct the reasoning that led him to leave his decision to chance) or in cases of mental illness (although a psychiatrist could partially decipher the reasoning at work), human decision-making is normally the result of a deliberation in which arguments are weighed, preferences are compared, and expectations are advanced – within analytical frameworks that can, in large part, be reconstituted. One must assume, moreover, that the more a decision appears to have weighty consequences, the more the contemplation undertaken aims to be fully rational. Certainly, a decision can be absurd or unreasonable. Yet one can describe, in a logical manner, how one arrived at either type of decision. Rational choice theory thus has the great merit of reminding us that, for any kind of science, everything is intelligible at the end of the day – and that nothing is by nature unknowable. Analysis of a decision sequence and articulation of its successive steps definitely does not enable one to fully explain an action undertaken. Yet it does allow scholars to shed light on a clearly defined sequence, one that exists in almost all situations – and thus to pursue such a sequence with the aim of more comprehensive knowledge.

63Next, rational choice theory has the mischievous quality of beating its detractors at their own game. In France, rational choice theory has often been disseminated via scholarship critical of it. While most of the contributors to Rational Choice in Political Science make an effort to locate its pertinent aspects, the tone of the book remains markedly reserved about it. [106] And yet, no sociologist or political scientist who rejects rational choice theory would ever consider themselves irrational in their critical assessments or, moreover, in the ensemble of their scientific work. The fundamental value to which a scientist adheres is evidently cognitive rationality. Arguments are made, and must be made, according to rigorous norms of reasoning. The logical soundness of an argument lies at the very basis of any scientific exchange. Is it not then curious that the political scientist can, in a single stroke, be an eminently rational actor who employs a logical series of arguments in scientific debate as well as someone who uses that rationality to demonstrate that rational choice theory cannot be accepted? A bit more modesty towards other social actors, [107] and a bit less forgetting of the principles upon which the political scientist’s own practices are founded, would not be unwelcome.

64Finally, by dismissing with excessive vigor the sociological theory of rational choice, do we not risk jeopardizing rationality as a possible foundation for political society itself? During the past few decades, philosophers, along with sociologists, political scientists, and scientists, have been searching for a universal ethic capable of ending the terrible state of conflict in the world. They believe they have found such a possibility in man’s use of reason – since all men are endowed with reason and able to make arguments. [108] While I am unable to develop this point here, I wish to refer briefly to the debates around what is termed “discourse ethics”. Who cannot see that a complete rejection of rational choice theory as a scientific paradigm would undermine all hope for the rationalization – and thus the pacification – of the world? This is not to say that I accept rational choice theory as my own, with its precise technical modalities that are outlined and examined in this article. It seems to me, however, that rational choice theory is indispensable for a more general taking into account of the place of rationality in thought. [109]